Enhanced recovery protocols in patients undergoing pancreatic surgery: An umbrella review

Abstract Aim To identify, synthesize and appraise the systematic reviews of ERAS for patients undergoing pancreatic surgery and facilitate ERAS implementation. Design An umbrella review was used to identify systematic reviews. Methods A systematic search following the PRISMA guidelines was used to search databases including PubMed, Embase, Cochrane Library, CINAHL, CNKI, WanFang and VJIP. AMSTAR 2 was used to appraise the quality of included reviews. Results Ten systematic reviews were included. The quality of all included systematic reviews was rated as “critically low.” The most frequently reported ERAS elements were epidurals analgesia/PCA (9/10), goal‐directed mobilization (9/10) and early removal of drains (9/10). Only one review mentioned audit protocol compliance. None of the included reviews reported discharge standards. Ten reviews reported decreased length of stay, seven reviews reported lower hospital costs, and six reviews reported decreased total complications rate. There were no adverse effects reported.


| INTRODUC TI ON
Enhanced recovery after surgery (ERAS), also known as "fast track," "critical pathways" or "clinical pathways" (Coolsen et al., 2013), is a multimodal, multidisciplinary approach to perioperative care which was introduced to improve patient outcomes and reduce healthcare costs (Ljungqvist et al., 2017). ERAS protocols have developed rapidly in the last decade and focussed on bundle elements including, but not limited to, preoperative information and teaching, decreased of stress, pain relief, early mobilization and early oral diet (Bond-Smith et al., 2016;Ljungqvist et al., 2017). ERAS guidelines have been developed for elective rectal/pelvic surgery (Nygren et al., 2012), radical cystectomy for bladder cancer (Cerantola et al., 2013), oesophagectomy (Findlay et al., 2014) and gastrectomy (Mortensen et al., 2014), and all suggest that ERAS protocols can decrease the length of hospital stay and reduce costs in these surgery procedures (Cerantola et al., 2013;Findlay et al., 2014;Mortensen et al., 2014;Nygren et al., 2012). Enhanced recovery after pancreatic surgery was developed cautiously because of the complexity of pancreatic surgery. Kennedy et al., (2007) initially reported the enhanced recovery protocol after pancreatic surgery in 2007, which reported improved patient outcomes. Coolsen et al., (2015) also reported that ERAS protocols were feasible and did not compromise patient outcomes in older adults. The first guideline for perioperative care for pancreaticoduodenectomy published in 2012 contained 27 recommendations (Lassen et al., 2012). Subsequently, further reviews on enhanced recovery after pancreatic surgery have shown that it is effective in decreasing hospital length of stay (Pecorelli et al., 2016), was less expensive and although no difference in readmission and postoperative morbidity was observed (Barton, 2016;Elhassan et al., 2019;Feng et al., 2017;Kagedan et al., 2015;Pecorelli et al., 2016;Perinel & Adham, 2016;Xie et al., 2016;Ypsilantis & Praseedom, 2009). However, existing reviews (Coolsen et al., 2013;Kagedan et al., 2015;Xie et al., 2016;Ypsilantis & Praseedom, 2009) demonstrated variability in the ERAS elements. Outcome indicators are inconsistently reported, which lead to uncertainty for clinical decision makers, and made clinical application challenging. The aim of this umbrella review was to identify, synthesize and appraise the systematic reviews on enhanced recovery after pancreatic surgery and provide critical analysis of ERAS interventions and outcome measures, and to facilitate ERAS implementation.

| Design
This is an umbrella review. The screening process followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline (Moher et al., 2009). The articles were identified from target databases according to the search strategy and were exported into Endnote X9 (Clarivate Analytics, Philadelphia PA). Duplicates were removed. The title and abstract of all articles were initially screened according to inclusion and exclusion criteria; full text screening against the inclusion and exclusion criteria followed.

| Quality appraisal
The AMSTAR 2 tool (Shea et al., 2017), which contains 16 discrete evaluation questions and was designed for critical appraisal of systematic reviews, was used to guide quality assessment. Seven of the AMSTAR 2 items are described as critical domains (Item 2,4,7,9,11,13,15); the remaining nine items are considered non-critical. Quality assessment is described as "high" (one or fewer non-critical weakness), "moderate" (more than one non-critical weakness), "low" (one critical flaw with or without non-critical weaknesses) and "critically low" (more than one critical flaw with or without non-critical weakness). Two authors independently undertook a quality assessment of included articles; a third author assisted with consensus moderation where required.

| Data abstraction and synthesis
Data extraction was independently conducted by two authors. An overview of study characteristics is provided in Table 2, ERAS elements of studies included in each review are summarized in Table 3, and patient outcomes are summarized Table 4.

| The characteristics of included reviews
A total of 10 systematic reviews met the inclusion criteria ( Figure 1).
Of the 10 systematic reviews included, seven reviews were from China (three were published in Chinese, and four were published in English) (Cao et al., 2019;Chen et al., 2019;Ji et al.,  there was one review each from UK, Netherlands and Canada, respectively. Two systematic reviews did not include a meta-analysis (Kagedan et al., 2015Ypsilantis & Praseedom, 2009). Only three reviews included randomized controlled trials (Chen et al., 2019;Lei et al., 2015;Zhang et al., 2018) (Table 2).

| The quality of included reviews
Quality appraisal using AMSTAR2 (Shea et al., 2017) identified that all included reviews had more than one item identified as a critical weakness; thus, the quality of all the reviews was deemed as "critically low" (Table S1). Critical domains of AMSTAR2 items 2 (Did the report of the review contain an explicit statement that the review methods were established prior to the conduct of the review and did the report justify any significant deviations from the protocol?) and item 13 (Did the review authors account for Risk of Bias in primary studies when interpreting/discussing the results of the review?) were not present for any included review.

| The ERAS elements and outcome indicators reported
The ERAS elements mentioned in the included reviews varied widely ( Table 3)

| Overview
ERAS is a patient-centred, evidence-based approach to patient care, which includes multiple interventions designed to facilitate optimal postoperative recovery (Ljungqvist et al., 2017). ERAS in pancreatic surgery has been reported in the literature for over a decade (Kennedy et al., 2007), with the first ERAS guideline for pancreatic surgery published in 2012 (Lassen et al., 2012), and the most recent update published in 2020 (Melloul et al., 2020). Compared with other subspecialties in general surgery, ERAS uptake in pancreatic surgery has been comparatively cautious (Lei et al., 2015). This may be attributed to the fact that pancreatic surgery is a complex, highly difficult procedure, and often associated with more postoperative complications (Lei et al., 2015).
In this umbrella review, we found that 9/10 reviews published after the original ERAS guideline for pancreatic surgery (Lassen et al., 2012) were made available, this shows that the release of the guidelines has promoted the application of ERAS in patients after pancreatic surgery. Only three reviews included randomized controlled trials, and this indicates that research in this field is still

TA B L E 3 Summary of ERAS elements in included reviews
8 7 9 6 7 5 1

Incisional wound infection
Pulmonary infection lacking high-quality evidence. Also, there may be bias on patient grouping, which made the results biased, decreasing the confidence of the conclusions of individual reviews.

| The quality of included reviews
In this umbrella review, we found that the overall quality of the 10 included reviews was "critically low" according to AMSTAR 2 criteria (Shea et al., 2017). For item 2, no review published their protocol in advance, which may introduce publication bias, and unsatisfactory indicators may be ignored and reported, such as adverse events and unexpected withdrawal from the trial. Having an "a priori" review protocol available prior to undertaking the review is recommended (Shamseer et al., 2016) and can increase the transparency of the review process. Item 13 is another critical domain of AMSTAR 2, where no authors accounted for risk of bias in individual studies when discussing the results of the review. Risk of bias in individual studies could lead to biases on evidence synthesis in the reviews.
If the conclusions of the review were applied to all patients after pancreatic surgery without analysis, it may endanger the safety of patients. Thus, these reviews do not provide the accurate and comprehensive synthesis of the available evidence (Shea et al., 2017).
Clinicians should be cautious in adopting the conclusions of these systematic reviews.
A less frequently reported effectiveness indicator was the impact of ERAS on the patient recovery process. For example, in only two reviews (Cao et al., 2019;Chen et al., 2019) was a decrease in the time of first flatulence reported, and time to first oral intake or mobilization was not reported in any review. It is suggested that more attention should be paid to observation of and reporting outcomes

Eligibility
Included reflecting functional status (e.g. physical activity, activities of daily living) and overall health (e.g. quality of life) of patient recovery when ERAS is used (Feldman et al., 2015).
Therefore, it is suggested that long-term follow-up should be used to observe the complications in order to ensure the safety of ERAS after pancreatectomy.

| Selection of ERAS elements
Some patient-related outcomes reported in these reviews were not always included, such as delayed gastric emptying (Cao et al., 2019;Coolsen et al., 2013;Ji et al., 2018;Xie et al., 2016;Xiong et al., 2016;Zhang et al., 2018) and pancreatic fistula (Cao et al., 2019;Ji et al., 2018;Xie et al., 2016;Xiong et al., 2016;Zhang et al., 2018). This may relate to the lack of consistency in which ERAS elements were implemented in the protocols evaluated in the original articles included in the 10 systematic reviews. The ERAS components most commonly included were the use of epidurals analgesia/PCA (Cao et al., 2019;Chen et al., 2019;Coolsen et al., 2013;Ji et al., 2018;Kagedan et al., 2015;Lei et al., 2015;Xie et al., 2016;Xiong et al., 2016;Ypsilantis & Praseedom, 2009), goal-directed mobilization (Cao et al., 2019;Chen et al., 2019;Coolsen et al., 2013;Ji et al., 2018;Lei et al., 2015;Xie et al., 2016;Xiong et al., 2016;Ypsilantis & Praseedom, 2009;Zhang et al., 2018) and early removal of drains (Chen et al., 2019;Coolsen et al., 2013;Ji et al., 2018;Kagedan et al., 2015;Lei et al., 2015;Xie et al., 2016;Xiong et al., 2016;Ypsilantis & Praseedom, 2009;Zhang et al., 2018). Most of the components were related to the postoperative phase, which was essential to promote optimal postoperative recovery. Among them, sufficient multimodal postoperative analgesia is strongly recommended as a core component for rapid recovery (Mendes et al., 2018;Partelli et al., 2016;Pecorelli et al., 2016), which has been found to reduce stress response to surgery and improving compliance to early goal-directed mobilization. It can also bring in the recovery of gastrointestinal function and early removal of the nasogastric drain (Williamsson et al., 2015), thus speed up the postoperative recovery process. In addition, some items of ERAS were already supported by clear high-level evidence and have a higher level of recommendation in the latest guideline, such as avoiding hypothermia, use of wound catheter, antimicrobial and thromboprophylaxis, preoperative nutritional interventions for patients with severe weight loss (Melloul et al., 2020), so priority should be given to these ERAS measures.
A protocol compliance audit was only included in one review as an ERAS intervention. (Coolsen et al., 2013) despite auditing and feedback of the ERAS protocol being identified as effective strategies in implementation studies (Bisch et al., 2018;Nelson et al., 2016), to help decrease ERAS practice variation. The latest guidelines published in 2020 also recommended that compliance should be documented as part of future trials (Melloul et al., 2020).
Protocol compliance was found to be low in the postoperative phase (Maessen et al., 2007), and auditing adherence to the protocol is recommended as a standard item in ERAS implementation (Pecorelli et al., 2017). Brown and Xhaja (2018) proposed that ERAS coordinators could be well positioned to audit the ERAS process and promote protocol adherence through weekly chart audits, particularly in the early stages of implementation. These suggestions are important given it has been shown that maintaining high compliance, sustainability and improving patient outcomes is possible when ERAS guidelines are implemented (Arrick et al., 2019;Pisarska et al., 2018).

| Limitations
There are some limitations in this umbrella review. First, we did not conduct a meta-analysis in this review, due to heterogeneity on patients' outcomes and unclear definition of outcomes in included systematic reviews. Secondly, because of the inconsistencies in the components implemented/reported in the included systematic reviews makes it difficult to clearly recommend which individual interventions are likely to be most beneficial. Last, because the quality of all the included reviews was low, it was difficult to draw clear conclusions to evaluate the benefit of the bundle of ERAS intervention.

| CON CLUS ION
In summary, the quality of existing systematic reviews on ERAS after pancreatic surgery suggests the further high-quality research in this area is required. In addition, the quality of systematic reviews can be improved to make available higher-quality evidence syntheses.
The feasibility and effectiveness of individual ERAS elements also require further investigation so that those with the highest-quality evidence is prioritized for implementation. Future research should incorporate economic evaluation, process evaluation and intervention compliance.

ACK N OWLED G EM ENT
The authors would like to thank Bonnie Dixon from Griffith University, Australia, and Guanhua Ren from Peking University First Hospital for their contribution to developing and undertaking the search of the literature.

CO N FLI C T O F I NTE R E S T
The authors had no conflict of interest to declare.

AUTH O R CO NTR I B UTI O N
All authors contributed to the plan and methodological study discussion of the review. Jing Li and Shuhui Yu: Database search, data extraction and quality assessment. Jing Li: Writing. Andrea P.
Marshall, Frances Lin and Shuhui Yu: Revision and approval of the final version.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.